The present invention relates to an image forming apparatus such as PPC and FAX and more particularly to an image magnification changing apparatus which reduces or enlarges the magnification of an image by image processing.
Conventionally, there has been widely used an image reducing method which stores image data necessary for reduction of image in multiple line memories and achieves reduction processing by a processing through a projection method.
Because, according to the prior art, the capacity of a necessary line memory increases as the reduction rate decreases (as the image size is reduced), the image processing apparatus needs to have a line memory having a capacity necessary for its minimum reduction rate. For example, in case of a reduction by 16%, a line memory for seven lines is required thereby leading to a rise of cost.
A conventional image enlarging method, as disclosed in, for example, Jpn. Pat. Appln. KOKAI Publication No. 5-7293, stores an inputted image in its page memory and executes enlargement processing in a sub-scanning direction while reading an image from the page memory. This prevents an operating speed of the ASIC from rising, and consequently, an image output synchronous with a printer synchronous signal is realized.
However, in case of such prior art, image data for one page needs to be stored, so that a memory amount increases tremendously thereby leading to an increase of cost. Further, when the magnification changing processing for reduction or enlargement is carried out, the density of a character edge portion drops so that a character or a narrow line becomes blurred.
Furthermore, when the enlargement processing and the filter processing are carried out on the same image, the filter size is increased exponential-functionally depending on the enlargement rate of the filter size. For example, if the filter size at 100% is set to N×N, in order to achieve the same characteristic as the filter characteristic at the time of 100% on an image after the enlargement, the filter size needs to be (N× magnification)×(N× magnification). More specifically, if both main scanning and sub-scanning are executed at 200%, N is 11, and 23×23 size is necessary. When the enlargement and filter processing are executed on an image, the filter size is increased, thereby leading to a rise of H/W (hard wear) cost, which is a problem to be solved.
To the contrary, when the filter processing is carried out before the enlargement and then, the enlargement processing is carried out, there occurs such a problem that the character and line edge become blurred, thereby sharpness thereof being dropped. The same thing can be said of image reduction processing.